EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS

Vol. 3, No. 1, 2010, 118-127

ISSN 1307-5543 www.ejpam.com

An Invitation to Proofs Without Words

Claudi Alsina1 and Roger B. Nelsen2
1
2

Universitat Politecnica de Catalunya, Secci de Matemtiques i Informtica, Barcelona, Spain

Lewis & Clark College, Department of Mathematical Sciences, Portland Oregon, USA

Abstract. Proofs without words are pictures or diagrams that help the reader see why a particular
mathematical statement may be true, and also see how one might begin to go about proving it true.
In some instances a proof without words may include an equation or two to guide the reader, but the
emphasis is clearly on providing visual clues to stimulate mathematical thought. While proofs without
words can be employed in many areas of mathematics (geometry, number theory, trigonometry, calculus, inequalities, and so on) in our invitation we examine only one area: elementary combinatorics.
In this article we use combinatorial proof methods based on two simple counting principles (the Fubini principle and the Cantor principle) to wordlessly prove several simple theorems about the natural
numbers.
2000 Mathematics Subject Classifications: 00A05
Key Words and Phrases: proofs without words, visual proofs, visualization in mathematics

1. Introduction
What are proofs without words? As you will see from this article, the question does not
have a simple, concise answer. Generally, proofs without words are pictures or diagrams that
help the reader see why a particular mathematical statement may be true, and also to see how
one might begin to go about proving it true. As Yuri Ivanovich Manin said, A good proof is
one that makes us wiser, a sentiment echoed by Andrew Gleason: Proofs really arent there
to convince you that something is true - theyre there to show you why it is true.
Proofs without words (PWWs) are regular features in two journals published by the Mathematical Association of America. PWWs began to appear in Mathematics Magazine about
1975, and in the College Mathematics Journal about ten years later. Many of these appear in
two collections of PWWs published by the Mathematical Association of America [8, 9].

Corresponding author.

Email address: laudio.alsinaup .edu (C. Alsina), nelsen lark.edu (R. Nelsen)
http://www.ejpam.com

118

c 2009 EJPAM All rights reserved.

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

119

But PWWs are not recent innovations-they have been around for a very long time, perhaps
first appearing in ancient Greece and China, and later in tenth century Arabia and renaissance
Italy. Today PWWs regularly appear in journals published around the world and on the World
Wide Web.
Some argue that PWWs are not really proofs, nor, for that matter, are they without
words, on account of equations which often accompany a PWW. Martin Gardner, in his popular Mathematical Games column in the October 1973 issue of Scientific American, discussed
PWWs as look - see diagrams. He said in many cases a dull proof can be supplemented
by a geometric analog so simple and beautiful that the truth of a theorem is almost seen at
a glance. It is in that spirit that we write this invitation to PWWs. In some instances we
include an equation or two to guide the reader, but the emphasis is clearly on providing visual
clues to stimulate mathematical thought. We encourage the reader to think about how the
picture proves the theorem before reading on. However, in each case we have included a
short description of what we hope the reader sees in each picture.
We believe there is a role for PWWs in mathematics classrooms from elementary schools
to universities. The ability to visualize is essential for success in mathematics, and George
Plyas Draw a figure. . . is classic pedagogical advice.

2. Combinatorial Proofs
PWWs can be employed in many areas of mathematics, to prove theorems in geometry,
number theory, trigonometry, calculus, inequalities, and so on. In our invitation we will
examine only one area: elementary combinatorics. In many theorems concerning the natural
numbers {1, 2, . . .}, insight can be gained by representing the numbers as sets of objects. Since
the particular choice of object is unimportant, in PWWs we usually use dots, squares, balls,
cubes, and other easily drawn objects.
In this article we will use combinatorial proof methods based on two simple counting
principles that can be applied to representations of natural numbers by sets of objects. The
principles are:
1. If you count the objects in a set in two different ways, you will obtain the same result;
and
2. If two sets are in one-to-one correspondence, then they have the same number of elements.
The first principle has been called the Fubini principle [11] after the theorem in multivariable calculus concerning exchanging the order of integration in iterated integrals. We call the
second the Cantor principle, after Georg Cantor (1845 - 1918), who used it extensively in his
investigations into the cardinality of infinite sets. The two proof techniques are also known as
the double - counting method and the bijection method, respectively.

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

120

3. Figurate Numbers
The idea of representing a number by a set of objects (perhaps as pebbles on the beach)
dates back at least to the ancient Greeks. When that representation takes the shape of a polygon such as a triangle or a square, the number is often called a figurate number. We begin
with some theorems and proofs about the simplest figurate numbers: triangular numbers and
squares.
Nearly every biography of the great mathematician Carl Friedrich Gauss (1777 - 1855)
relates the following story. When Gauss was about ten years old, his arithmetic teacher asked
the students in class to compute the sum 1 + 2 + 3 + . . . + 100, anticipating this would keep
the students busy for some time. He barely finished stating the problem when young Carl
came forward and placed his slate on the teachers desk, void of calculation, with the correct
answer: 5050. When asked to explain, Gauss admitted he recognized the pattern 1 + 100 =
101, 2 + 99 = 101, 3 + 98 = 101, and so on to 50 + 51 = 101. Since there are fifty such pairs,
the sum must be 50 101 = 5050. The pattern for the sum (adding the largest number to
the smallest, the second largest to the second smallest, and so on) is illustrated in Figure 1,
where the rows of balls represent positive integers.

Figure 1
The number t n = 1 + 2 + 3 + . . . + n for a positive integer n is called the nth triangular
number, from the pattern of the dots on the left in Figure 1. Young Carl correctly computed
t 100 = 5050. However, this solution works only for n even, so we first prove
Theorem 1. For all n 1, t n =

n(n+1)
.
2

Proof. See Figure 2.

Figure 2

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

121

We arrange two copies of t n to form a rectangular array of balls in n rows and n + 1

columns. Then we have 2t n = n(n + 1), or t n = n(n + 1)/2.
The counting procedure in the preceding combinatorial proof is the Fubini principle. We
employ the same procedure to prove that sums of odd numbers are squares.
Theorem 2. For all n 1, 1 + 3 + 5 + . . . + (2n 1) = n2 .
Proof. We give two combinatorial proofs in Figure 3.

(a)

(b)

Figure 3
In Figure 3a, we count the balls in two ways, first as a square array of balls, and then by
the number of balls in each L-shaped region of similarly colored balls (the Fubini principle). In
Figure 3b, we see a one-to one correspondence (illustrated by the color of the balls) between
a triangular array of balls in rows with 1, 3, 5, . . . , (2n 1) balls, and a square array of balls
(the Cantor principle).
The same idea can be employed in three dimensions to establish the following sequence
of identities:
1 + 2 = 3,
4 + 5 + 6 = 7 + 8,
9 + 10 + 11 + 12 = 13 + 14 + 15, etc.
Note that each row begins with a square number. The general pattern
n2 + (n2 + 1) + . . . + (n2 + n) = (n2 + n + 1) + . . . + (n2 + 2n)
can be proved by induction, but the following visual proof is much nicer.

Figure 4

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

122

In Figure 4, we see the n = 4 version of the identity where counting the number of small
cubes in the pile in two different ways yields 16 + 17 + 18 + 19 + 20 = 21 + 22 + 23 + 24.
There are many nice relationships between triangular and square numbers. The simplest
is perhaps the one illustrated in the right side of Figure 3b: t n1 + t n = n2 . Two more are
given in the following lemma (setting t 0 = 0 for convenience):
Lemma 1. For all n 0, (a) 8t n = 1 = (2n + 1)2 , and (b) 9t n + 1 = t 3n+1 .
Proof. See Figure 5 (where we have replaced balls by squares).
(a)

(b)

Figure 5

Lemma 1 enables us to prove the following two theorems.

Theorem 3. There are infinitely many numbers that are simultaneously square and triangular.
Proof. Observe that
8t n (8t n + 1)

= 4t n (2n + 1)2 ,
2
so if t n is square, then so is t 8t n . Since t 1 = 1, this relation generates an infinite sequence
of square triangular numbers, e.g., t 8 = 62 and t 288 = 2042. However, there are square
triangular numbers such as t 49 = 352 and t 1681 = 11892 that are not in this sequence.
t 8t n =

Theorem 4. Sums of powers of 9 are triangular numbers, i.e., for all n 0,

1 + 9 + 92 + . . . + 9n = t 1+3+32 +...+3n
Proof. See Figure 6.

Figure 6

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

123

As a consequence, in base 9 the numbers 1, 11, 111, 1111, . . . are all triangular.
The next theorem presents a companion to the identity t n1 + t n = n2 .
Theorem 5. The sum of the squares of consecutive triangular numbers is a triangular number,
i.e., t 2n1 + t 2n = t n2 for all n 1.
Proof. See Figure 7, where we illustrate the square of a triangular number as a triangular
array of triangular numbers.

Figure 7
You may have noticed that the nth triangular number is a binomial coefficient, i.e., t n =
One explanation for this is that each is equal to n(n + 1)/2, but this answer sheds little
light on why it is true. Here is a better explanation using the Cantor principle:
n+1

.
2

Theorem 6. There exists a one-to-one correspondence between a set of t n objects and the set of
two-element subsets of a set with n + 1 objects.
k

Proof. See Figure 8 [6], and recall that the binomial coefficient 2 is the number of
ways to choose 2 elements from a set of k elements. The arrows denote the correspondence
between an element of the set with t n elements and a pair of elements from a set of n + 1
elements.

...

1
2
3

n
n+1

Figure 8

4. Sums of Squares, Triangular Numbers, and Cubes

Having examined triangular numbers and squares as sums of integers and sums of odd
integers, we now consider sums of triangular numbers and sums of squares.

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

124

Theorem 7. For all n 1, 12 + 22 + 32 + . . . + n2 = n(n + 1)(2n + 1)/6.

Proof. We give two proofs. The first is in Figure 9.

AAA
AAA
AA
AAA
AAA
AA
AA AAA
AAA
AAA
AA
AAA
AAA
AA
AA
AAA
AAA
AAA
AAA
AA
Figure 9

We exhibit a one-to-one correspondence between three copies of 12 + 22 + 32 + . . . + n2

and a rectangle whose dimensions are 2n + 1 and 1 + 2 + . . . + n = n(n + 1)/2 [4]. Hence
3(12 + 22 + 32 + . . . + n2 ) = (2n + 1)(1 + 2 + . . . + n) from which the result now follows.
For a second proof, see Figure 10 [5].

... n-1
n-1 n-1
n n n ... n
n

...

n n-1 ... 4 3 2
... 3 2 1
n n-1

n
n-1 n
n-2 n-1 n

...

n
n n-1
n n-1 n-2

2
1

3
2

4 ... n-1 n
... n-1 n

2n+1
2n+1 2n+1
2n+1 2n+1 2n+1

...

...

1
2 2
3 3 3

2n+1 2n+1 ... 2n+1

...
2n+1 2n+1
2n+1

Figure 10
Here we write each square k2 as a sum of k ks, then place those numbers in a triangular
array, create two more arrays by rotating the triangular array by 120 and 240 , and add
corresponding entries in each triangular array.
Theorem 8. For all n 1, t 1 + t 2 + t 3 + . . . + t n = n(n + 1)(n + 2)/6.
Proof. See Figure 11.

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

125

(n+1)

Figure 11
Here we stack layers of unit cubes to represent the triangular numbers. The sum of the
triangular numbers is total number of cubes, which is the same as the total volume of the
cubes. To compute the volume, we slice off small pyramids (shaded gray) and place each
small pyramid on the top of the cube from which it came. The result is a large right triangular
pyramid minus some smaller right triangular pyramids along one edge of the base. Thus
t1 + t2 + t3 + . . . + t n =

1
6

(n + 1)3 (n + 1)

1
6

n(n + 1)(n + 2)
6

In the above proof we evaluated the sum of the first n triangular numbers by computing volumes of pyramids. This is actually an extension of the Fubini principle from simple
enumeration of objects to additive measures such as length, area and volume. The volume
version of the Fubini principle is: computing the volume of an object in two different ways yields
the same number; and similarly for length and area. We cannot, however, extend the Cantor
principle to additive measures - for example, one can construct a one-to-one correspondence
between the points on two line segments with different lengths.
Theorem 9. For all n 1, 13 + 23 + 33 + . . . + n3 = (1 + 2 + 3 + . . . + n)2 = t 2n .
Proof. Again, we give two proofs. In the first, we represent k3 as k copies of a square with
area k2 to establish the identity [3, 7].

Figure 12
In Figure 12, we have 4(13 + 23 + 33 + . . . + n3 ) = [n(n + 1)]2 (for n = 4).
For the second proof, we use the fact that 1+2+3+. . .+(n1)+n+(n1)+. . .+2+1 = n2
(we leave it as an exercise for the reader to draw a picture with balls in a square array and

C. Alsina and R. Nelsen / Eur. J. Pure Appl. Math, 3 (2010), 118-127

126

count the balls by diagonals in the square array) and consider a square array of numbers
(rather than balls) in which the element in row i and column j is i j, and sum the numbers in
two different ways. See Figure 13 [10].

1
2

2
4

3
6

...
...

n
2n

1
2

2
4

3
6

...
...

n
2n

...

3n

...

3n

n2

n 2n 3n

...

...

...

...

...
...

...

...
...

n 2n 3n

...

n2

Figure 13
Summing by columns yields
n
X
i=1

i+2

n
X

i
i=1

+ ... + n

n
X

i
i=1

!2

n
X

i=1

while summing by the L-shaped shaded regions yields 1 12 + 2 22 + . . . + n n2 + =

Pn

i=1 i

We conclude this section with a theorem representing a cube as a double sum of integers.
Pn Pn
Theorem 10. For all n 1, i=1 j=1 (i + j + 1) = n3 .
Proof. We represent the double sum as a collection of unit cubes and compute the volume
of a rectangular box composed of two copies of the collection. See Figure 14.

Figure 14
Pn Pn
Observe that two copies of the sum S = i=1 j=1 (i + j +1) fit into a rectangular box with
base n2 and height 2n, hence computing the volume of the box in two ways yields 2S = 2n3 ,
or S = n3 .

5. Conclusion
In this short survey we have looked at visual proofs using representations of numbers by
sets of objects. Of course, there are many other ways to represent numbers (not necessarily

REFERENCES

127

natural numbers), including representing them as lengths of segments, areas of plane figures,
volumes of objects. With such representations, other techniques can be employed in the
proofs, such as rotations, translations, reflections, and other transformations that preserve
length, area, or volume. For a comprehensive survey of these techniques (and many others),
see the books Math Made Visual: Creating Images for Understanding Mathematics and When
Less Is More: Visualizing Basic Inequalities [1, 2].

References
[1] C Alsina and R Nelsen. Math Made Visual: Creating Images for Understanding Mathematics. Mathematical Association of America, Washington, 2006.
[2] C Alsina and R Nelsen. When Less Is More: Visualizing Basic Inequalities. Mathematical
Association of America, Washington, 2009.
[3] A Cupillari. Proof without words. Mathematics Magazine, (62):259, 1989.
[4] M Gardner. Mathematical games. Scientific American, page 115, October 1973.
[5] S Kung. Sum of squares. College Mathematics Journal, (20):205, 1989.
[6] L Larson. A discrete look at 1 + 2 + . . . + n. College Mathematics Journal, (16):369382,
1985.
[7] W Lushbaugh. (no title). Mathematical Gazette, (49):200, 1965.
[8] R Nelsen. Proofs Without Words: Exercises in Visual Thinking. Mathematical Association
of America, Washington, 1993.
[9] R Nelsen. Proofs Without Words II: More Exercises in Visual Thinking. Mathematical
Association of America, Washington, 2000.
[10] F Pouryoussefi. Proof without words. Mathematics Magazine, (62):323, 1989.
[11] S Stein. Existence out of chaos. In R Honsberger, editor, Mathematical Plums, pages
6293. Mathematical Association of America, Washington, 1979.